May 15, 2020

Are you planning Technical Debt in Sprint Planning?

Scrum Events by Agilemania

Scrum is a straightforward framework to develop complex software products in a complex environment but hard to practice. But why? I am not going in detail of why, but wanted to highlight one problem with the development team as per my understanding.

The development team means a cross-functional and self-organizing team of developers. But who is the developer? Isn't it everyone who contributes to developing software? A coder, tester, dba, UI designer, content writer, and business analyst? Now, look at your team. Is it cross-functional? Yes, you have,  because you have technicians for all skills, but are those technicians are developers? Maybe not. They are just coders, a tester, a dba, and UI designers, and are waiting for their skill-related tasks. They are not working as developers intending to develop software, but they are doing their tasks based on their skills.

 

Key things that will help you to understand the characteristics of the development team. 

Assume L&D (Learning and Development Dept) is about to organize a few training and announcing like below.

1st announcement "we are planning to have a workshop for Scrum Developer, so send your nomination." You will find that only coders have subscribed to this workshop.

Another announcement, "we are organizing Agile Testing workshop, so send your nomination." This time only tester will subscribe. The same happens with Agile Business Analysis and UI/UX training. 

What reflects above? The team is still divided based on skills coding, analysis, UI/UX, and testing. Tester feels everything related to testing is their domain, and similarly, coders think all coding work is part of their work.

In reality, testers write code to test production codes, and coders write tests to perform unit and integration tests of code.

 

Let's look at it differently.

During sprint planning team has prepared below sprint backlog. 

 

Sprint Board

 

What it reflects? There are tasks, but those tasks are skills-based and not component-based. Isn't it a wrong way of creating sub-tasks? If not? Then read the below situation.

Assume Alex (coder) and Martha (tester) working on these three stories.

Alex completed the coding part of story one in 2 days, and Martha has started testing, so what Alex will do next?

Most likely, Alex will start working on the coding part of story 2. But is it useful? Why can't Alex pick up other work from the same story? Or why Martha was waiting for Alex to complete coding to start testing? Why can't she pair with Alex to test in parallel or at least started testing APIs?

Many people asked me what's wrong if Alex starts working on the coding part of story 2. Let's see the below situation.

Alex completed coding of 1st story and started coding 2nd story. Martha found a bug in story 1, so what does she have to do? She reached out to Alex and wanted to discuss it, but Alex is busy and asked her to come later.

What will Martha do next? Most likely, she will open a bug tracking tool and log a newly found bug over there. Is it good? No, but why Not? I will write about bug logging challenges separately, but focus on other aspects of it for now.

Alex is done with the coding of Story 2 and started looking into bugs that were logged by Martha. He identified the root cause of a bug and framework level code changes needed to fix it. But that will also impact the code that he has just completed for story 2.

What will Alex do? If he changes whatever is needed to fix at the root, it will generate rework and delay delivery. Another option is a shortcut to fix the bug (patchwork), so it should not impact the 2nd story code.

The majority of time coders choose 2nd option because coders are very emotional about their code (just kidding). There are multiple reasons for selecting the 2nd option, but the result will be "Technical Debt."

What else?

It seems you missed another big issue that I have highlighted in this video, so recognize and think twice why it is essential to avoid skill-based task creation. 

What can we do to avoid such a scenario?

1st don't create sub-tasks like above but try to divide into components. The best is not to create sub-tasks and keep stories small. Work in the pairing as a unit instead of a passing story from one to another. The team should focus on finishing one that already in progress rather than starting a new.

Thanks for reading! If you have any queries, then reach out to me. I teach and coach the same in my Professional Scrum Developer training.